Desktop version of Pytorch memory viz, handling profiles of 10GB and more, with quality-of-life improvements

Access large language models from the command-line

Aidan Bench attempts to measure <big_model_smell> in LLMs.

The AdEMAMix Optimizer: Better, Faster, Older.

Code for `BCD Nets: Scalable Variational Approaches for Bayesian Causal Discovery`, Neurips 2021

🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.

QLoRA: Efficient Finetuning of Quantized LLMs

[ICML 2023] Reflected Diffusion Models (https://arxiv.org/abs/2304.04740)

The agent engineering platform.

JAX Synergistic Memory Inspector

Curation of prompts that are known to be adversarial to large language models

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

Causal Analysis of Agent Behavior for AI Safety

This repository contains the source code of the EMNLP 2020 paper Interactive Fiction Game Playing as Multi-Paragraph Reading Comprehension with Reinforcement Learning.

Numerical differential equation solvers in JAX. Autodifferentiable and GPU-capable. https://docs.kidger.site/diffrax/

Knowledge-Aware RL agents with Commonsense Reasoning

A gym interface for AI safety gridworlds created in pycolab.

Checking D-separations and I-equivalence in Bayesian Networks.

Compressed numerical arrays that support high-speed random access

Pretrained deep learning models for Jax/Flax: StyleGAN2, GPT2, VGG, ResNet, etc.

Official code for Score-Based Generative Modeling through Stochastic Differential Equations (ICLR 2021, Oral)

Code for "Infinitely Deep Bayesian Neural Networks with Stochastic Differential Equations"

Implementation of Unconstrained Monotonic Neural Network and the related experiments. These architectures are particularly useful for modelling monotonic transformations in normalizing flows.

Python Black for Emacs

Syntax highlighting and other enhancements for the Julia REPL

JAX (Flax) implementation of algorithms for Deep Reinforcement Learning with continuous action spaces.

Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more

Normalizing Flows using JAX

Benchmarking Linear Assignment Problem Solvers